pounds in 1901, 1169 pounds in 1902, and 1589 pounds in 1903; 

 and Hall's book contains the following: 



"In March, 1897, and in July, 1899, all the plants were removed by hand, burnt 

 and their ashes returned, and the soil was carefully picked over by hand for 

 the Sclerotia of the fungus, Sclerotinia trifoliorum, many of which were found. 

 The soil was also dressed with carbon bisulfid as a fungicide, before fresh 

 seed was sown. In 1903, which was a favorable year for the growth of clover, 

 a fair plant was obtained by reseeding, and in the spring of 1904 the best crop 

 for many years was cut from this plot." 



Director Hall expresses the opinion that the fungus named is 

 not the only cause of " clover sickness." 



Finally, it should be understood that, while the Rothamsted 

 field experiments have been conducted with extreme care, there 

 are some possible sources of error, and the Rothamsted Station 

 has been very careful to point these out where they are of probable 

 consequence. Warrington, in his Rothamsted lectures (Bulletin 

 No. 8, Office of Experiment Stations, United States Department 

 of Agriculture) , delivered before the Association of American Agri- 

 cultural Colleges and Experiment Stations, in 1891, under the pro- 

 visions of the Lawes Agricultural Trust, makes the following 

 statements : 



"The earlier experimental fields at Rothamsted were not arranged as skill- 

 fully as the later ones; thus, Broadbalk wheat field has long, narrow plots, 

 and the influence of the manure of neighboring plots is in some cases distinctly 

 felt. The barley experiments in Hoos field are the best laid out ; here the plots 

 are nearly square; they have each an area of one fifth of an acre." 



(In the author's opinion, tenth-acre plots, 2 by 8 rods or i by 

 16 rods, or fifth-acre plots, 4 by 8 rods or 2 by 1 6 rods, are more 

 satisfactory than square plots for field experiments, because greater 

 uniformity between plots is thus secured; but in all cases a pro- 

 tecting border of at least one fourth rod should completely surround 

 every plot, the same crops being grown upon the border as upon 

 the plot proper. This requires a half-rod division strip between 

 plots, and wherever needed, an additional uncultivated strip of 

 grass sod should be left between the plots.) 



On the Grass Park at Rothamsted an imaginary line is the only 

 division between the plots, but the ground is never broken, and the 

 fertilizers are applied as top dressings with exactness (a cloth screen 



